Microbes Environ. Vol. 28, No. 3, 370–380, 2013

https://www.jstage.jst.go.jp/browse/jsme2 doi:10.1264/jsme2.ME13030

Seasonal Transition of Active Bacterial and Archaeal Communities in

Relation to Water Management in Paddy Soils

HIDEOMI ITOH1†*, SATOSHI ISHII1††, YUTAKA SHIRATORI2, KENSHIRO OSHIMA3, SHIGETO OTSUKA

1,

MASAHIRA HATTORI3, and KEISHI SENOO1

1Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, The University of

Tokyo, 1–1–1 Yayoi, Bunkyo-ku, Tokyo 113–8657, Japan; 2Niigata Agricultural Research Institute, 857 Nagakuramati,

Nagaoka, Niigata 940–0826, Japan; and 3Department of Computational Biology, Graduate School of Frontier Sciences,

The University of Tokyo, 5–1–5 Kashiwanoha, Kashiwa, Chiba 277–8561, Japan

(Received March 5, 2013—Accepted June 19, 2013—Published online September 5, 2013)

Paddy soils have an environment in which waterlogging and drainage occur during the rice growing season.Fingerprinting analysis based on soil RNA indicated that active microbial populations changed in response to watermanagement conditions, although the fundamental microbial community was stable as assessed by DNA-basedfingerprinting analysis. Comparative clone library analysis based on bacterial and archaeal 16S rRNAs (5,277 and5,436 clones, respectively) revealed stable and variable members under waterlogged or drained conditions. Clonesrelated to the class Deltaproteobacteria and phylum Euryarchaeota were most frequently obtained from the samplescollected under both waterlogged and drained conditions. Clones related to syntrophic hydrogen-producing bacteria,hydrogenotrophic methanogenic archaea, rice cluster III, V, and IV, and uncultured crenarchaeotal group 1.2 appearedin greater proportion in the samples collected under waterlogged conditions than in those collected under drainedconditions, while clones belonging to rice cluster VI related to ammonia-oxidizing archaea (AOA) appeared at higherfrequency in the samples collected under drained conditions than in those collected under waterlogged conditions.These results suggested that hydrogenotrophic methanogenesis may become active under waterlogged conditions,whereas ammonia oxidation may progress by rice cluster VI becoming active under drained conditions in the paddy field.

Key words: Paddy soil, Soil microbial diversity, Clone libraries

Unlike upland agricultural soils, paddy soils have unique

environmental conditions as a result of water management

practices (25). In Japan and many other countries, paddy

fields are drained completely before rice cultivation and are

waterlogged temporally during the rice growing season. As

the time of harvest approaches, paddy fields are drained again

and maintained in the same condition until the next cultivation

season. Such water management practices greatly impact the

soil biogeochemical properties of paddy fields. The soil

environment is oxic during drained periods, whereas this

environment becomes anoxic during waterlogged periods.

Various anaerobic biochemical processes occur during water-

logged periods, including nitrate, metal, and SO42− reductions,

and methanogenesis (25, 29). Although seasonal transition

of the soil geochemical properties of paddy fields have been

studied in detail (25), responses of microbial communities to

environmental changes in paddy soils are not well understood.

Microbial community shifts have been examined in Italian

paddy fields throughout the rice cultivation period (26). It

was revealed that the composition of the methanogenic

archaeal community was constant during the cultivation

period despite seasonal transition of methane production

activity. However, these studies were conducted on the basis

of the 16S rRNA gene derived from DNA extracted from

soils; therefore, the transition of active microbial communities

under field conditions has remained unclear. DNA-based gene

analysis may detect not only viable but also dead or quiescent

cells (19, 36). In contrast, rRNA-based analysis is believed

to detect metabolically active populations because growing

cells contain more ribosomes and rRNA is degraded imme-

diately after the suspension of metabolic activity (39, 22).

Although some microbes are known to maintain the amount

of rRNA under starvation conditions (9, 10), previous studies

demonstrated that community shifts were more pronounced

when identified with an RNA-based approach than with

DNA-based analysis (7, 45, 51). Thus, RNA-based analysis

may be more suitable to investigate the responses of microbial

communities to environmental changes.

RNA-based studies were conducted to investigate bacterial

or methanogenic archaeal community shifts in relation to

flooded conditions in paddy soil microcosms (38, 35) and

Japanese paddy fields (23, 56), indicating some active

members among limited groups, bacteria or methanogenic

archaea. Flooded paddy fields are well known major

environmental sources of methane production, gasses con-

tributing to the greenhouse effect (31). Many previous studies

focused on methanogenic archaeal communities in rice paddy

soil (12, 33, 26, 57). However, little is known about the

* Corresponding author. E-mail: hideomi-itou@aist.go.jp /

hitoh19j@gmail.com;

Tel: +81–11–857–8909; Fax: +81–11–857–8915.† Present address: National Institute of Advanced Industrial Science

and Technology (AIST) Hokkaido Center, 2–17–2–1, Tsukisamu-

higashi, Toyohira-ku, Sapporo, Hokkaido 062–8517, Japan†† Present address: Division of Environmental Engineering, Faculty

of Engineering, Hokkaido University, North 13, West 8, Kita-ku,

Sapporo, Hokkaido 060–8628, Japan

Active Microbes in Paddy Soil 371

transition of whole prokaryotic communities because simul-

taneous assessment of bacterial and archaeal communities

was not performed in any previous field studies based on soil

RNA.

Sampling frequency in previous studies may also have

limited their assessment of microbial responses to changes

in soil geochemical properties (23). Lüdemann et al. (32)

showed that oxygen was depleted in paddy soil 2.2 mm below

the surface after 7 days of incubation. Because the anoxic

condition stimulates the anaerobic processes involved in soil

geochemical properties, frequent sampling (e.g., weekly) is

necessary to assess the responses of active microbial

populations to changes in soil geochemical properties.

Consequently, our first objective was to investigate the

seasonal transition of both bacterial and archaeal communities

in the paddy field in relation to water management during

cultivation seasons. On a weekly or biweekly basis, we

performed soil sampling and simultaneous assessment of

bacterial and archaeal communities by quantitative PCR

(qPCR) and fingerprinting analysis using DNA and RNA in

the soil samples collected from the paddy fields. The second

objective was to identify the active microbial population

under waterlogged and drained conditions. Comparative

analysis of RNA-based clone libraries was performed to

identify the active microbes responding to changes in the soil

condition.

Materials and Methods

Study site and soil sampling

The paddy field used in this study is located at Niigata AgriculturalResearch Institute (Nagaoka, Niigata, Japan; 37°26'N, 138°52'E).The soil type is classified as Gley Lowland soil. Rice (Oryza sativaL., cv. Koshihikari) has been cultivated in the field since 2003 asa single summer crop from April to September. Water managementstages were divided into five categories: before waterlogging (stageBW; until April 29), waterlogging (stage W; April 30–June 18),temporal drainage (stage T; June 19–30), intermittent drainage withcycles of artificial drainage and irrigation (stage I; July 1–August31), and after complete drainage (stage CD; after September 1).Details of site management in 2009 are described in the supplemen-tary information (Fig. S1).

Soil samples were collected at 20 time points from April toOctober 2009 (Fig. S1). At each sampling event, 10 soil cores(3 cm in diameter) were collected from the plow layer at a depthof 0 to 10 cm after removing the surface water and mixed well ina plastic bag. Part of the composite sample was immediately frozenin liquid nitrogen, transported with dry ice, and stored at −80°Cuntil used for extraction of nucleic acid. The remaining soil sampleswere maintained at 4°C until they were used for the analysis of soilcharacteristics as described below.

Soil characteristics

Soil moisture and temperature were monitored at a depth of5 cm in the experimental field every 2 h during the samplingperiod using an Em5b Analog Data Logger (Decagon Devices,Pullman, WA, USA) equipped with an EC5 soil moisture sensorand ECT temperature sensor (Decagon Devices). Soil Eh wasmeasured at a depth of 5 cm on all sampling dates using threereplicate platinum-tipped electrodes and Eh indicator PRN-41(Fujiwara Scientific, Tokyo, Japan) in the field. N2O and CH4 fluxin the field were measured using the closed chamber method (37).Soil pH and NH4-N, NO3-N, NO2-N, Fe2+, Mn2+, and SO4

2−

concentrations were measured as described previously (41, 17).Measurement of Fe2+ concentration and extraction of Mn2+ were

performed on the day of sampling to minimize oxidation of thereduced metals. Denitrification and nitrification activity weremeasured as described in the supporting information.

Nucleotide preparation

From 1 g (wet weight) of each soil sample, RNA was extractedusing the RNA PowerSoil Total RNA Isolation Kit (MoBioLaboratories, Solana Beach, CA, USA) according to the manufac-turer’s instructions. DNA was extracted simultaneously through theRNA extraction step using the RNA PowerSoil DNA ElutionAccessory Kit (MoBio Laboratories). Crude RNA was purified usingthe Turbo DNA-free Kit (Applied Biosystems, Foster City, CA,USA) and RNA Clean-Up Kit-5 (Zymo Research, Irvine, CA, USA)to remove DNA and PCR inhibitors as much as possible. First-strand complementary DNA (cDNA) was synthesized by incubatingtotal RNA (100 ng) in 20 µl of reaction mixture composed ofRandom primer 6 mer (Takara Bio, Otsu, Japan), 5× PrimeScriptbuffer, SUPERase-In RNase inhibitor (20 units) (Ambion, Austin,TX, USA), and PrimeScript reverse transcriptase (200 units) (TakaraBio) at 42°C for 60 min. The concentration and integrity of theprepared nucleotide solution were determined by spectrophotometryusing the NanoDrop ND-1000 spectrophotometer (NanoDropTechnologies, Wilmington, DE, USA), electrophoresis on 1.5%agarose gel stained with ethidium bromide, and qPCR with standardaddition. Three separate nucleotide extraction were performed fromthe original composite soil samples made from 10 soil cores. Thereproducibility of replicate nucleotide extractions was confirmed tobe similar by denaturing gradient gel electrophoresis (DGGE) (Fig.S2). Three replicates of DNA and cDNA for each soil sample wereused for qPCR. Pooled DNA and cDNA from three replicateextractions were used for PCR-DGGE and clone library analysesas described below.

qPCR

qPCR was performed to amplify bacterial and archaeal 16S rRNAgenes using Power SYBR Green PCR Master Mix (AppliedBiosystems) and the StepOne system (Applied Biosystems). Thereaction mixture comprised 2× SYBR Green PCR Master Mix, 0.2µM forward and reverse primer pairs ((357F and 520R for bacteria(41); A364a and A934b for archaea (15), 0.5 µg/µl BSA, and DNAor cDNA as a template. The PCR conditions were as follows:initial denaturation at 95°C for 10 min, followed by 40 cycles of95°C for 30 s, 58°C for 30 s, and 72°C for 30 s (for bacteria) or40 cycles of 94°C for 30 s, 66.5°C for 30 s, and 72°C for 50 s(for archaea). The amount of archaeal or bacterial 16S rRNA andthe 16S rRNA gene copies was calculated on the basis of thestandard curve constructed using the dilution series of the plasmidsolutions inserted with near full-length 16S rRNA gene sequencesfrom Methanosarcina barkeri NBRC 100474 (NR_074253) orCupriavidus metallidurans JCM 21315T (NR_074704), respectively.

PCR-DGGE

PCR amplification and DGGE were performed as describedpreviously (41, 15). PCR product size was confirmed by electro-phoresis on 1.5% agarose gel stained with ethidium bromide, andamplicons were purified using the Wizard SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA). Bacterial DGGE bandsunique to the sample collected on May 21 were excised from thegel, and DNA was eluted and sequenced as described previously(41).

Clone library construction and sequencing

For clone library analysis, one DNA sample obtained from thesoil sample collected on June 10, 2009, and four cDNA samplesobtained from the soil sample collected on April 23, June 10 and18, and October 1, 2009 were used. Bacterial and archaeal partial16S rRNAs were amplified under PCR conditions with the optimizedPCR cycle numbers, as described in the supporting information.PCR products were verified and purified as described above and

ITOH et al.372

were cloned into the pCR TOPO vector using the TOPO TA CloningKit (Invitrogen) as described by the manufacturer’s instruction. Asa result, 10 clone libraries (bacterial and archaeal 16S rRNA genesequence libraries from five nucleotide samples) were constructed.From each library, >1000 colonies were randomly picked using theGeneTac G3 Picking System (DIGLAB, Holliston, MA, USA), andtheir insert sequences were amplified using PCR-1 (5'-GTGCTGCAAGGCGATTAAGTTGG-3') and PCR-2 (5'-TCCGGCTCGTATGTTGTGTGGA-3') primers. PCR amplicons were treated withexonuclease I and shrimp alkaline phosphatase (GE Healthcare,Uppsala, Sweden) and sequenced using a BigDye Terminator v3.1Kit with an automated ABI 3730xl capillary sequencer (AppliedBiosystems). Bacterial fragments were sequenced from both endsof DNA using primers M13F and M13R (17). High quality sequencedata derived from the two primers were assembled, and the consensussequences were obtained using the Phred/Phrap program (8). Singlestrand sequencing of archaeal samples was performed using primerM13.

Phylogenetic analysis

Chimeric sequences from these 16S rRNA clone libraries weredetected using the Mallard program with a 99.9% cutoff (1) andremoved from the libraries. The resulting sequences were processedfor taxonomic assignment at the family and higher levels using RDPClassifier program ver. 2.2 (55) with an 80% confidence threshold.Archaeal or bacterial nucleotide sequences from the five clonelibraries were aligned using ClustalW2 (28), and communitysimilarity among the libraries was examined using the Fast UniFracweb interface (13). Rarefaction curves, Chao1 and ACE richnessindices, and Shannon and Simpson’s diversity indices werecalculated using Mothur program ver. 1.14.0 (47). Sequences fromeach library were also clustered into OTUs with 3% differencesusing the Mothur program. OTUs specific to each library and sharedOTUs between two libraries were identified using the templatematch method with a similarity index of <0.10 as a cutoff value(16, 17) using R software ver. 2.10.1. (http://www.r-project.org/).Phylogenetic tree was constructed using the neighbor-joiningmethod with bootstrap test (1,000 replicates) using MEGA ver. 4.0.2(53).

Statistical analysis

Analysis of variance (ANOVA) was performed using R softwarever. 2.10.1 to analyze differences in soil characteristics and qPCRdata among samples. R software was also used to perform theprincipal component analysis (PCA) of the DGGE profiles digitizedusing CS Analyzer 3 (ATTO, Tokyo, Japan) (16) and Fisher’s exacttest to analyze taxonomic distributions among the clone libraries.

Nucleotide sequence accession number

The nucleotide sequences reported in this paper were depositedin the DDBJ/Genbank/EBI databases under accession numbersAB661339–AB661347 (DGGE bands) and AB650607–AB661319(sequences obtained from the clone libraries).

Results

Seasonal transition of soil biogeochemical properties

During the sampling period, we measured 14 soil bio-

geochemical properties. Volumetric water content in paddy

soil increased after waterlogging, reached a plateau (>50%)

during waterlogged periods (stage W; April 30–June 18), and

decreased with water drainage (Fig. 1A). Soil Eh decreased

gradually after waterlogging and was maintained at approx-

imately −230 mV at a later stage W (June 4–18; Fig. 1B).

Soil Eh increased in response to water drainage and was

maintained at >+350 mV after complete drainage (stage CD;

after September 1) (Fig. 1B). NH4+ concentration increased

after fertilizer application on April 23 and decreased thereafter

(Fig. 1C). NO3− concentration was low before waterlogging

(stage BW; until April 29) and became <1 mg N kg−1 soil

by 2 weeks after waterlogging (on April 30; Fig. 1C). Nitrite

was not detected (<1 mg NO2−-N kg−1 soil) during the

sampling period (data not shown). Potential nitrification

activity increased in the initial stage W (April 30–May 21)

and in stage CD (Fig. 1D). Potential denitrifying activity

changed in accordance with soil Eh (Fig. 1D). Similarly,

concentrations of Mn2+ and Fe2+ and CH4 flux increased and

SO42− concentration decreased in accordance with the changes

in soil Eh (Figs. 1E and 1F). N2O flux from the paddy field

was low (<0.1 µg N2O-N m−2 h−1) throughout the sampling

period (data not shown). Fig. 1G shows the results of qPCR

analysis performed using RNA extracted from soils. The

amounts of bacterial 16S rRNA were similar among stage

BW, W, T, and I with no significant different (P>0.05).

However, they decreased from stage W to stage CD, and the

amounts during stages W and CD were significantly different

(P<0.05). A similar tendency was also observed during

analysis using archaeal 16S rRNA (Fig. 1G). In contrast,

qPCR analysis using DNA extracted from soils showed that

the amount of bacterial and archaeal 16S rRNA genes did

not change throughout the sampling period (Fig. 1H).

Seasonal transition of the structure of bacterial and archaeal

communities throughout the cultivation period

Based on the DGGE profiles targeting bacterial 16S rRNA,

the bacterial community structure appeared to be similar

among samples (Fig. S3A); however, soil samples in each

water management stage were grouped together while

examining the band intensities (Fig. 2A). The DGGE profiles

of samples in stages BW and CD, in which the soils were

both under drained conditions, were close together on the

PCA plot (Fig. 2A). The DGGE profile of the sample collected

on May 21 was distinct from that of the other samples (Figs.

2A and S3A). Bands unique to this profile were excised and

sequenced. Phylogenetic analysis of the sequences from seven

major bands showed that the sequences of band A belonged

to the phylum Betaproteobacteria (Table S1) and those of

bands B to G were related to the phylum Cyanobacteria, not

chloroplasts (Table S1 and Fig. S4). Similar to the DGGE

profiles targeting bacterial 16S rRNA, DGGE profiles

targeting archaeal 16S rRNA were similar across samples

(Fig. S3B). However, PCA analysis of band intensities

showed that the archaeal DGGE profiles changed with time

(Fig. 2B). In addition to the analysis based on RNA, we

performed DGGE analysis on the basis of DNA extracted

from soils. In contrast to the results of RNA-based analysis,

bacterial and archaeal community structures were stable over

time (Figs. S3C and S3D), and no apparent groupings of

samples were observed on the PCA plots (Figs. 2C and 2D).

Comparative clone library analysis

qPCR and DGGE analyses indicated that active populations

in bacterial and archaeal communities were different between

stages W and CD (Fig. 2). On the basis of these data, we

selected four RNA samples (D1R at stage BW, W1R and

W2R at stage W, and D2R at stage CD) derived from soil

samples collected on April 23, June 10, June 18, and October

Active Microbes in Paddy Soil 373

1, respectively, for further analysis. In addition, one DNA

sample, W1D in stage W, was used for further analysis. We

performed comparative clone library analysis to identify

active populations in bacterial and archaeal communities and

their seasonal dynamics in detail. The Mallard program (1)

detected 2.3 to 5.7% and 0.6 to 3.7% of the sequences as

chimeras from bacterial and archaeal clone libraries, respec-

tively. After removal of these chimeric sequences, we

obtained 5,277 and 5,436 sequences from bacterial and

archaeal clone libraries, respectively (Table 1), and used them

for subsequent analysis. The diversity of communities derived

from each library was analyzed using the Mothur program

(47). At 3% difference, 692 to 731 and 132 to 191 operational

taxonomic units (OTUs) were detected from bacterial and

Fig. 1. Seasonal transition of soil biogeochemical properties in paddy soils for the sampling period. (A) Daily average soil water content (blackline) and soil temperature (gray line) at 5-cm depth in the experimental paddy field; (B) soil redox potential (Eh) (closed circles) and pH (opencircles) at 5-cm depth; (C) concentrations of NH4

+ (closed triangles) and NO3− (open triangles); (D) nitrification activity (closed reverse triangles)

and denitrification activity (open reverse triangles); (E) concentrations of Fe2+ (closed squares) and Mn2+ (open squares); (F) concentrations ofSO4

2− (closed diamonds) and CH4 flux from the experimental field (open diamonds); (G) amount of bacterial (smaller closed circles) and archaeal(smaller open circles) 16S rRNAs measured by RNA-based qPCR; and (H) copy number of bacterial (smaller closed circles) and archaeal (smalleropen circles) 16S rRNA genes measured by DNA-based qPCR. Mean ±SD is shown (n=3). Samples as indicated with arrows and dotted lines wereused for clone library analysis. Letters shown at the bottom indicate the water management stages: before waterlogging (stage BW; until 29 April),waterlogging (stage W; April 30–June 18), temporal drainage (stage T; June 19–30), intermittent drainage with cycles of artificial drainage andirrigation (stage I; July 1–August 31), and after complete drainage (stage CD; after September 1) (see Fig. S1).

ITOH et al.374

archaeal libraries, respectively (Table 1). The Chao1 and

ACE values as well as Shannon and Simpson diversity indices

were higher in bacterial libraries than in archaeal libraries

(Table 1). Library coverage (Cx) values ranged from 43 to

54% and 88 to 93% in bacterial and archaeal libraries,

respectively. Similarities among libraries analyzed by

weighted Fast UniFrac (13) suggested that libraries con-

structed from DNA were distinguished from those constructed

from RNA (Figs. 3A and 3B), regardless of the target

molecule (i.e., bacterial or archaeal 16S rRNA). AW1R and

AW2R libraries were very close to each other (Fig. 3B). The

difference between AW1R and AW2R was not significant

based on Fast UniFrac P test with 500 permutations (P=0.2),

while differences among the other bacterial and archaeal

libraries were significant (P<0.05).

Fig. 4A shows the phylum- or class-level distribution of

clones obtained from the libraries targeting bacterial 16S

rRNA. Deltaproteobacteria dominated among bacteria in all

four samples derived from RNA. The relative abundances of

the phylum Cyanobacteria and Bacteroidetes seemed to

increase in the BW1R and BW2R libraries constructed using

soil samples at stage W than in the BD1R and BD2R libraries

constructed using soil samples at stage BW and CD, although

with no significant difference (P>0.05) based on Fisher’s

exact test. Within the deltaproteobacterial clones, clones

belonging to the order Myxococcales were dominant, ranging

from 35.6 to 42.5% (Fig. 4B); however, their proportions

Fig. 2. Transition of bacterial and archaeal community structuresas assessed by principal component analysis (PCA) based on thedenaturing gradient gel electrophoresis (DGGE) profiles. PCA plotsbased on (A) RNA-based DGGE targeting bacterial 16S rRNA; (B)RNA-based DGGE targeting archaeal 16S rRNA; (C) DNA-basedDGGE targeting bacterial 16S rRNA genes; and (D) DNA-basedDGGE targeting archaeal 16S rRNA genes are shown. Legend:closed circle, stage BW; closed triangles, stage W; closed square, stageT; closed diamonds, stage I; open circles, stage CD. Arrows indicatethe temporal transition sequences. Values in parentheses show thepercentage of community variation explained by each component. Thesymbol with asterisk in panel A indicates the profiles derived from thesample collected on May 21 (see Supplementary information).

Table 1. Diversity properties of bacterial and archaeal clone libraries

TargetLibrary

IDNo. of

sequencesaNo. of

OTUsbCxc

Estimated OTUs richnessd Diversity indices

Chao1 Ace Shannone 1/Simpson

Bacteria BW1D 999 708 0.43 2623 (2169, 3216) 5719 (5084, 6447) 6.33±0.06 571

BD1R 1069 693 0.52 1843 (1580, 2183) 3382 (3034, 3782) 6.30±0.06 569

BW1R 1058 731 0.48 1990 (1712, 2349) 3539 (3167, 3967) 6.42±0.05 859

BW2R 1059 725 0.48 2115 (1802, 2519) 3914 (3493, 4398) 6.39±0.06 738

BD2R 1092 692 0.54 1802 (1547, 2134) 2960 (2660, 3305) 6.31±0.06 622

Archaea AW1D 1085 134 0.93 341 (236, 553) 528 (440, 643) 3.92±0.07 32

AD1R 1063 191 0.88 857 (544, 1444) 1539 (1328, 1790) 4.04±0.09 28

AW1R 1075 189 0.89 581 (408, 892) 1078 (923, 1267) 4.24±0.08 42

AW2R 1107 176 0.92 463 (332, 706) 647 (551, 770) 4.23±0.07 43

AD2R 1106 132 0.93 572 (332, 1101) 731 (613, 879) 3.62±0.09 19

a Number of sequences after removal of chimeric sequences by Mallard.b Calculated with Mothur at the 3% distance level.c Calculated from the equation, Cx=1−(n/N), where ‘n’ is the number of OTUs composed of only one sequence (singleton) and N is the total number

of sequences.d Numbers in parentheses are 95% confidence intervals.e Mean ±95% confidence intervals.

Fig. 3. Principal coordinate analysis (PCoA) plots based on (A) bac-terial and (B) archaeal clone libraries. PCoA plots were generatedusing Fast UniFrac analysis based on the weighted algorithm with nor-malization. Values in parentheses show the percentage of communityvariation explained by each coordinate. Samples in circles have signifi-cant similarity (P>0.05). The first letter of library ID (A or B) repre-sents the target community: B, bacteria; A, archaea. The final letter oflibrary ID (D or R) represents the nucleic acid used for analysis: D,DNA; R, RNA. Libraries with ID containing “D1,” “W1,” “W2,” and“D2” were derived from soil samples collected on Apr. 23, June 10,June 18, and October 1, respectively. So, D1 and D2 libraries werederived from drained conditions, while W1 and W2 libraries werederived from waterlogged conditions.

Active Microbes in Paddy Soil 375

did not change with library type. In contrast, proportions of

the orders Desulfobacterales, Syntrophobacterales, and the

family Syntrophorhabdaceae varied with library type and

were greater (P<0.05) in the libraries constructed using

soil samples in stage W than in those constructed using

other samples (Fig. 4B). Fig. 4C shows the phylum-level

distribution of clones obtained from libraries targeting

archaeal 16S rRNA. Euryarchaeota dominated throughout

the sampling period with their proportions ranging from 65.5

to 76.7% (Fig. 4C). Most (90.9 to 95.1%) of the euryarchaeal

sequences belonged to the class Methanomicrobia, which

includes the orders Methanocellales, Methanomicrobiales,

and Methanosarcinales. The relative abundances of Metha-

nocellales and Methanomicrobiales were 1.9 to 2.3 times

higher (P<0.05) in the libraries constructed using soils under

waterlogged conditions (AW1R and AW2R) than in those

constructed using soils under drained conditions (AD1R and

AD2R) (Fig. 4D), respectively. Because we constructed clone

libraries using DNA and RNA extracted from the same soil

samples, we could compare the taxonomic distributions of

clones between the two libraries (W1D and W1R). Fig. 4

shows that the taxonomic distributions were distinct from

each other, similar to the results of Fast UniFrac analysis

described above. More detailed taxonomic compositions are

shown in Tables S3 and S4.

Identification of active populations in bacterial and archaeal

communities in paddy soils under waterlogged or drained

conditions

The results of DGGE and clone library analyses indicated

that active populations in bacterial and archaeal communities

in paddy soils may have changed in response to water

management conditions. To identify the microbes responsive

to waterlogged or drained soil conditions, we performed

OTU-based analysis. Using the Mothur program and the

template match method (16, 17), we identified OTUs specific

to the libraries (W1R and W2R) that were constructed using

soils under waterlogged conditions (OTUw) and those

Fig. 4. Relative distributions of the clones at different taxonomic resolutions. Resolutions at (A) bacterial phylum and proteobacterial class; (B)deltaproteobacterial order; (C) archaeal phylum; and (D) Methanomicrobia order and Methanosarcinales family are shown. The order name for thefamily Syntrophorhabdaceae is unidentified (43). Sequences were classified using the RDP Classifier with a threshold level of 80%.

ITOH et al.376

specific to the libraries (D1R and D2R) that were constructed

using soils under drained conditions (OTUd). In the bacterial

clone libraries, 95 and 78 OTUs were identified as being

OTUw and OTUd (OTUwB and OTUdB), respectively. The

OTUwB was dominated by Deltaproteobacteria (26.3%),

Betaproteobacteria (14.7%), and Cyanobacteria (8.4%),

while the OTUdB was dominated by Deltaproteobacteria

(15.4%), Actinobacteria (15.4%), and Betaproteobacteria

(12.8%) (Table S2). Among Deltaproteobacteria, Desulfo-

bacterales, Syntrophobacterales, Syntrophorhabdaceae, and

the NRBW cluster (unidentified cluster detected frequently

in our experimental paddy soil) appeared only in OTUwB

(Fig. 5).

Similar analysis was performed with OTUs obtained from

the archaeal clone libraries. Forty-two and 21 OTUs were

identified as being OTUw and OTUd (OTUwA and OTUdA),

respectively. Among these, 67.2% and 57.1% of OTUwA

and OTUdA, respectively, were assigned as Euryarchaeota

by the Ribosomal Database Project (RDP) Classifier ver. 2.2

(55). Fig. 6A shows the phylogenetic tree constructed on the

basis of the representative sequences from OTUwA and

OTUdA belonging to Euryarchaeota. These OTUwA were

distributed across Methanocellales, rice cluster III and V, the

families Methanomicrobiaceae, Methanosaetaceae, and the

Fig. 5. Phylogenetic relatedness of the deltaproteobacterial 16S rRNA recovered from representative sequences from specific OTUs. SequencesIDs obtained in this study are indicated by colored smaller circles, black (OTUwB) or gray (OTUdB). Numbers in parentheses are the numbers ofclones in each OTUs that originated from the four libraries (BD1R, BW1R, BW2R, and BD2R). The size of right larger circle indicates relativeabundance of the total number of clones belonging to each cluster and the circle colors indicate association to either OTUwB or OTUdB as well assequence ID. NRBW and NRB clusters indicate the unidentified clusters with less similarity to known sequences despite their frequent detection inour experimental paddy soil. Bootstrap values (>70%) with 1,000 replicates are shown next to the branches. The 16S rRNA sequence of Bacillussubtilis (NC_00964) was used as an outgroup.

Active Microbes in Paddy Soil 377

NRAW cluster (unidentified cluster detected frequently in

our experimental paddy soil). In contrast, OTUdA appeared

frequently in Methanosarcinaceae and Methanosaetaceae of

Methanosarcinales. In addition to Euryarchaeota, 32.8% and

42.9% of OTUwA and OTUdA, respectively, were assigned

as the phylum Crenarchaeota or unclassified archaea by the

RDP Classifier ver. 2.2. The phylogenetic relationship among

the representative sequences from OTUwA and OTUdA

belonging to Crenarchaeota or unclassified archaea was

also examined (Fig. 6B). These OTUwA were related to

uncultured Crenarchaeaota Group 1.2, Group 1.1a, and rice

cluster IV, whereas OTUd were related to Group 1.2, rice

cluster VI, Group1.1a, Group1.1c, and the NRAD cluster

(unidentified cluster detected frequently in our experimental

paddy soil). The majority (65.0%) of the sequences from

OTUwA were related to Group 1.2, whereas only 3.2% of

the sequences from OTUdA were related to this cluster. In

contrast, many (39.8%) of the sequences from OTUdA were

related to rice cluster VI.

Discussion

We showed the seasonal transition of soil biogeochemical

properties in our experimental paddy field throughout the

cultivation period (Fig. 1). During stage W, soil Eh decreased

gradually, and anaerobic processes such as denitrification,

metal and SO42− reductions, and methanogenesis progressed

sequentially, as also observed in previous studies (25, 59,

33). In contrast, these anaerobic processes were depressed

under drained conditions. Finally, soil biogeochemical prop-

erties of the samples in stage CD were close to those in stage

BW. The observed seasonal transition of soil characteristics

confirmed that the soil environment became aerobic in stages

BW and CD and anoxic at stage W, especially in the later

stage (June 4–18) (Fig. 1). Based on the seasonal transition

of the amount of 16S rRNA, both bacteria and archaea were

suggested to be more abundant in stage W than stage CD

(Fig. 1G). Similar to our results, Conrad and Klose also

showed that the copy numbers of bacterial and archaeal 16S

rRNA genes increased in paddy soil microcosms amended

with rice straw after waterlogging (6). Our experimental field

received rice straw in October 2008; however, most remained

undegraded in soil in spring before cultivation in 2009

because of the low temperature (<6°C, daily average air

temperature) in winter. Microbes could use the remaining

rice straw to support their growth in stage W under warm

Fig. 6. A. Phylogenetic relatedness of the euryarchaeotal 16S rRNA recovered from representative sequences from specific OTUs. Sequences IDobtained in this study are colored black (OTUwA) or gray (OTUdA). Numbers in parentheses are the numbers of clones in each OTUs originatingfrom the four libraries (AD1R, AW1R, AW2R, and AD2R). Circle size indicates the relative abundance of the total number of clones belonging toeach cluster, and the circle colors indicate the deviation of OTUwA or OTUdA as well as sequence ID. NRAW cluster indicates the unidentifiedcluster with less similarity to known sequences despite frequent detection under waterlogged conditions in our experimental paddy soil. Bootstrapvalues (>70%) with 1,000 replicates are shown next to the branches. The 16S rRNA sequence of Bacillus subtilis (NC_00964) was used as anoutgroup. B. Phylogenetic relatedness of crenarchaeal and unclassified archaeal 16S rRNA recovered from representative sequences from specificOTUs. NRAD cluster indicates the unidentified cluster with less similarity to known sequences despite frequent detection under drained conditionsin our experimental paddy soil.

ITOH et al.378

condition (24). Under the drained condition in stage CD, the

decrease of prokaryotic 16S rRNA might be affected by a

longer drought period and lower temperature (Fig. 1G).

The results of RNA-based DGGE analysis and clone library

analysis suggested that active populations in bacterial and

archaeal communities responded to water management

conditions, although overall community structures were

stable over time as identified by DNA-based DGGE analysis

(Figs. 2 and 3). Previous field studies based on DNA or

phospholipids fatty acids directly extracted from soils did not

identify the effects of water management on the potential

structures of bacteria and methanogenic archaea in paddy

soils (23, 40, 56). These results suggested that RNA-based

analysis is more sensitive for detecting active populations in

the paddy field soil responsive to specific stimuli (e.g., water

management conditions). Indeed, some studies based on the

soil microcosm could detect the transitions of structure of

microbial communities in paddy soil by the DNA-based

method. However, they constructed the soil microcosm by

adding some carbon or nitrogen sources or waterlogged with

air-dried soil to trace the transition of structure of microbial

communities (17, 33). Compared with the natural conditions,

these stimulations might be more extreme, so changes to the

structure of microbial communities were suggested to be

detectable even through the DNA-based approach. Moreover,

the PCA plots of bacterial DGGE profiles based on RNA at

stage BW and CD were closer together than those of archaeal

DGGE profiles, suggesting that the structure of active

bacterial communities may recover more quickly from

waterlogging stimuli than those of archaeal communities in

the paddy field. Although their growth rate or resistance to

oxygen may affect the difference, further studies with

replicates are needed.

Comparative clone library analysis showed that clones

related to Deltaproteobacteria were most abundant in all

samples (Fig. 4A), in contrast to the bacterial community

structure in other soil environments (18, 49). Most

Deltaproteobacteria are known to be strict anaerobes,

except some groups related to Myxococcales (27).

Deltaproteobacteria were suggested to dominate in paddy

soil due to the temporal anoxic condition during the

waterlogged period. Methanosarcinales were also most

abundant in all samples (Fig. 4D), which are strict anaerobes

producing methane (5). Some members related to

Methanosarcinales appeared at high frequency despite the

oxic condition under drained conditions in stage BW and

CD, when methane flux was undetectable (Figs. 1F and 6A).

Other studies based on 16S rRNA genes and their transcripts

also reported that the structure of methanogenic archaea was

maintained during the cultivation season and became stable

even in the drained period when methane flux was not

detected (56, 35). Liu showed that methanogens were fragile

with oxygen in the medium but survived even under oxic

conditions if given the soil particles (30). Anaerobic

Deltaproteobacteria and methanogens were suggested to

survive (may be resting) even during the drained seasons in

the paddy soil, maintaining a certain amount of rRNA, as

reported in some bacteria (9, 10).

Within Deltaproteobacteria, clones related to Desulfo-

bacterales, Syntrophobacterales, and Syntrophorhabdaceae

increased under waterlogged conditions (Figs. 4B and 5).

Considering the increase of bacterial population size under

waterlogged conditions indicated by qPCR analysis, the

amount of these groups was suggested to increase under

waterlogged conditions. These bacteria can use SO42− or pro-

tons as electron acceptors (42, 43). Their clones were fre-

quently detected in other paddy soils (58, 34), suggesting

that these Deltaproteobacteria may play an important role

in SO42− reduction and hydrogen production under anoxic

conditions in paddy fields. The anoxic environment caused

by waterlogging may also develop an anaerobic archaeal

community in paddy soils. Although Methanosarcinales

clones were the most dominant in all soil samples, the

proportions of clones related to Methanomicrobiales and

Methanocellales increased in soils under anoxic conditions

(Figs. 4D and 6A). Because Methanomicrobiales and

Methanocellales archaea are known to be hydrogenotrophic

methanogens (11, 46), they most probably produce methane

from CO2 and H2 in paddy soils. Previous in vitro experi-

ments indicated that the occurrence of hydrogenotrophic

methanogenesis was supported by the presence of Syntro-

phobacterales and Syntrophorhabdaceae, which are H2

producers and require H2-utilizing organisms, such as meth-

anogens and SO42− reducers, as syntrophic partners (20, 43).

On June 10 and 18, SO42− concentration became undetectable

and methane flux increased greatly (Fig. 1F); therefore, H2

produced by Syntrophorhabdaceae and Syntrophobacterales

was most probably consumed by methanogenic archaea

rather than SO42−-reducing bacteria. A recent microcosm

study also showed the increase of the relative abundance of

hydrogenotrophic methanogens related to the family

Methanomicrobiaceae (order Methanomicrobiales) and

Methanocellaceae (order Methanocellales) in paddy soil

after waterlogging, based on terminal restriction fragment

length polymorphism analyses of mcrA transcripts encoding

the alpha subunit of methyl coenzyme M reductase involved

in methane production reaction (35). Combined with these

results, Methanomicrobiales and Methanocellales are sug-

gested to increase their cell numbers through methanogenesis

more rapidly than other methanogenic archaea. Ma et al.

reported that the Methanocellales organism has a unique set

of genes encoding antioxidant enzymes on its genome (35).

It is possible that Methanocellales have some tolerance to

oxygen, although the other methanogens do not, and could

grow advantageously under waterlogged conditions after the

drained period.

In addition to hydrogenotrophic methanogenic archaea,

rice clusters III, IV, and V were revealed to become more

active under waterlogged conditions. These clusters were

detected frequently in Italian and Asian paddy soils in

previous studies (33, 21, 44), although there is no information

about their seasonal transition. Although the ecological

function of these rice clusters remains unclear, these rice

clusters may prefer anoxic conditions and have anaerobic

respiration systems to survive under anoxic conditions in

paddy soils.

In contrast to soils under waterlogged conditions, clones

related to rice cluster VI (equivalent to Group 1.1b) appeared

at high frequency under drained conditions (Fig. 6B).

Considering the similar amount of archaea in stage BW and

Active Microbes in Paddy Soil 379

W (Fig. 1G), rice cluster VI was suggested to become active

under oxic conditions, before waterlogging at least. Repre-

sentative sequences belonging to rice cluster VI were closely

related to Candidatus Nitrososphaera sp., an ammonia-

oxidizing archaea (AOA) (14, 53). Recently, Group 1.1a and

1.1b containing AOA isolates were proposed to represent a

new archaeal cluster, the phylum Thaumarchaeota (2, 50).

Similar to our study, clones of ammonia monooxygenase

gene (amoA) for AOAs were also obtained from both

rhizosphere and bulk soils in paddy fields (3). In addition,

clones related to rice cluster VI were frequently retrieved

from paddy soils before waterlogged incubation in a micro-

cosm study (4). Rice cluster VI may become active under

drained conditions and play an important role in ammonia

oxidation in paddy soil.

This study showed the seasonal transition of active

populations in bacterial and archaeal communities in response

to marked changes in soil biogeochemical properties in paddy

fields. Simultaneous assessment of bacterial and archaeal

communities indicated that hydrogenotrophic methano-

genesis communities and rice clusters III, IV, and V became

active under waterlogged conditions, but they were depressed

under drained conditions in the paddy field. This study also

showed that rice cluster VI related to AOA tended to appear

at high frequency under drained conditions, whereas both

DNA- and RNA-based analysis of 16S rRNA indicated that

the fundamental microbial community was stable despite the

marked changes of soil geochemical properties. Although

RNA molecules are thought to be an indicator of active

microbes, some researchers have reported organisms main-

taining a constant amount of rRNA under starvation condi-

tions, as described above (8, 9). mRNA molecules could be

a stricter indicator of active microbes since the half-life of

mRNA is extremely short (48). Further studies with more

replicates and frequent sampling are needed, and metatran-

scriptomic approaches covering not only rRNA sequences

but also mRNA sequences may improve our understanding

of the seasonal transition of active microbes in paddy fields

(54).

Acknowledgements

We thank Dr. Masahito Hayatsu for technical advice aboutarchaeal DGGE analysis. This work was supported by (i) theProgram for the Promotion of Basic and Applied Researches forInnovations in Bio-oriented Industry, Tokyo, Japan, (ii) the projectfor Evaluation, Adaptation and Mitigation of Global Warming inAgriculture, Forestry and Fisheries (2006–2009) in the Ministry ofAgriculture, Forestry and Fisheries, Tokyo, Japan, and (iii) a Grant-in-Aid for Scientific Research (A) (22248038) and for a researchfellow for Scientific Research from the Japan Society for thePromotion of Science, Tokyo, Japan.

References

1. Ashelford, K.E., N.A. Chuzhanova, J.C. Fry, A.J. Jones, and A.J.Weightman. 2006. New screening software shows that most recentlarge 16S rRNA gene clone libraries contain chimeras. Appl. Environ.Microbiol. 72:5734–5741.

2. Brochier-Armanet, C., B. Boussau, S. Gribaldo, and P. Forterre. 2008.Mesophilic crenarchaeota, proposal for a third archaeal phylum, theThaumarchaeota. Nat. Rev. Microbiol. 6:245–252.

3. Chen, X.P., Y.G. Zhu, Y. Xia, J.P. Shen, and J.Z. He. 2008. Ammonia-oxidizing archaea, important players in paddy rhizosphere soil?Environ. Microbiol. 10:1978–1987.

4. Chin, K.J., T. Lukow, and R. Conrad. 1999. Effect of temperature onstructure and function of the methanogenic archaeal community in ananoxic rice field soil. Appl. Environ. Microbiol. 65:2341–2349.

5. Conrad, R. 1999. Contribution of hydrogen to methane productionand control of hydrogen concentrations in methanogenic soils andsediments. FEMS Microbiol. Ecol. 28:193–202.

6. Conrad, R., and M. Klose. 2006. Dynamics of the methanogenicarchaeal community in anoxic rice soil upon addition of straw. Eur. J.Soil Sci. 57:476–484.

7. Eichler, S., R. Christen, C. Höltje, P. Westphal, J. Bötel, I. Brettar,A. Mehling, and M.G. Höfle. 2006. Composition and dynamics ofbacterial communities of a drinking water supply system as assessedby RNA- and DNA-based 16S rRNA gene fingerprinting. Appl.Environ. Microbiol. 72:1858–1872.

8. Ewing, B., L. Hillier, M.C. Wendl, and P. Green. 1998. Base-callingof automated sequencer traces using phred. I. Accuracy assessment.Genome Res. 8:175–185.

9. Flärdh, K., P.S. Cohen, and S. Kjelleberg. 1992. Ribosomes existin large excess over the apparent demand for protein synthesisduring carbon starvation in marine Vibrio sp. strain-CCUG-15956. J.Bacteriol. 174:6780–6788.

10. Fukui, M., Y. Suwa, and Y. Urushigawa. 1996. High survival effi-ciency and ribosomal RNA decaying pattern of Desulfobacter latus, ahighly specific acetate-utilizing organism, during starvation. FEMSMicrobiol. Ecol. 19:17–25.

11. Garcia, J.L., B.K.C. Patel, and B. Ollivier. 2000. Taxonomic, phylo-genetic, and ecological diversity of methanogenic Archaea. Anaerobe6:205–226.

12. Großkopf, R., P.H. Janssen, and W. Liesack. 1998. Diversity andstructure of the methanogenic community in anoxic rice paddy soilmicrocosms as examined by cultivation and direct 16S rRNA genesequence retrieval. Appl. Environ. Microbiol. 64:960–969.

13. Hamady, M., C. Lozupone, and R. Knight. 2010. Fast UniFrac:Facilitating high-throughput phylogenetic analyses of microbial com-munities including analysis of pyrosequencing and PhyloChip data.ISME J. 4:17–27.

14. Hatzenpichler, R., E.V. Lebedeva, E. Spieck, K. Stoecker, A. Richter,H. Daims, and M. Wagner. 2008. A moderately thermophilic ammonia-oxidizing crenarchaeote from a hot spring. Proc. Natl. Acad. Sci.U.S.A. 105:2134–2139.

15. Hoshino, Y.T., S. Morimoto, M. Hayatsu, K. Nagaoka, C. Suzuki, T.Karasawa, M. Takenaka, and H. Akiyama. 2011. Effect of soil typeand fertilizer management on archaeal community in upland fieldsoils. Microbes Environ. 26:307–316.

16. Ishii, S., K. Kadota, and K. Senoo. 2009. Application of a clustering-based peak alignment algorithm to analyze various DNA finger-printing data. J Microbiol. Methods 78:344–350.

17. Ishii, S., M. Yamamoto, M. Kikuchi, K. Oshima, M. Hattori, S.Otsuka, and K. Senoo. 2009. Microbial populations responsive todenitrification-inducing conditions in paddy soil, as revealed bycomparative 16S rRNA gene analysis. Appl. Environ. Microbiol.75:7070–7078.

18. Janssen, P.H. 2006. Identifying the dominant soil bacteria taxa inlibraries of 16S rRNA and 16S rRNA genes. Appl. Environ. Micro-biol. 72:1719–1728.

19. Josephson, K.L., C.P. Gerba, and I.L. Pepper. 1993. Polymerase chainreaction detection of nonviable bacterial pathogens. Appl. Environ.Microbiol. 59:3513–3515.

20. Kato, S., and K. Watanabe. 2010. Ecological and evolutionaryinteractions in syntrophic methanogenic consortia. Microbes Environ.25:145–151.

21. Kemnitz, D., K.J. Chin, P. Bodelier, and R. Conrad. 2004. Commu-nity analysis of methanogenic archaea within a riparian floodinggradient. Environ. Microbiol. 6:449–461.

22. Kerkhof, L., and P. Kemp. 1999. Small ribosomal RNA content inmarine Proteobacteria during non-steady state growth. FEMSMicrobiol. Ecol. 30:253–260.

23. Kikuchi, H., T. Watanabe, Z. Jia, M. Kimura, and S. Asakawa. 2007.Molecular analyses reveal stability of bacterial communities in bulksoil of a Japanese paddy field: Estimation by denaturing gradientgel electrophoresis of 16S rRNA genes amplified from DNAaccompanied with RNA. Soil Sci. Plant Nutr. 53:448–458.

ITOH et al.380

24. Kimura, M., T. Minoda, and J. Murase. 1993. Water-soluble organicmaterials in paddy soil ecosystem, II. Effects of temperature oncontents of total organic materials, organic acids, and methane inleachate from submerged paddy soils amended with rice straw. SoilSci. Plant Nutr. 39:713–724.

25. Kimura, M. 2000. Anaerobic microbiology in waterlogged rice fields.p. 35–138. In J.M. Bollarg, and G. Stotzky (ed.), Soil Biochemistry,vol. 10. Marcel Dekker Inc., New York.

26. Krüger, M., P. Frenzel, D. Kemnitz, and R. Conrad. 2005. Activity,structure and dynamics of the methanogenic archaeal community in aflooded Italian rice field. FEMS Microbiol. Ecol. 51:323–331.

27. Kuever, J., F.A. Rainey, and F. Widdel, 2005. Class IV. Deltaproteo-bacteria class. nov. In D.J. Brenner, N.R. Krieg, and J.T. Staley, (eds),Bergey’s Manual of Systematic Bacteriology, 2nd edn. vol. 2. NewYork: Springer.

28. Larkin, M.A., G. Blackshields, N.P. Brown, et al. 2007. Clustal Wand Clustal X version 2.0. Bioinformatics 23:2947–2948.

29. Liesack, W., S. Schnell, and N.P. Revsbech. 2000. Microbiology offlooded rice paddies. FEMS Microbiol. Rev. 24:625–645.

30. Liu, C.T., T. Miyaki, T. Aono, and H. Oyaizu. 2008. Evaluation ofmethanogenic strains and their ability to endure aeration and waterstress. Curr. Microbiol. 56:214–218.

31. Lowe, D.C. 2006. Global change: A green source of surprise. Nature439:148–149.

32. Lüdemann, H., I. Arth, and W. Liesack. 2000. Spatial changes in thebacterial community structure along a vertical oxygen gradient inflooded paddy soil cores. Appl. Environ. Microbiol. 66:754–762.

33. Lueders, T., and M. Friedrich. 2000. Archaeal population dynamicsduring sequential reduction processes in rice field soil. Appl. Environ.Microbiol. 66:2732–2742.

34. Lueders, T., B. Pommerenke, and M.W. Friedrich. 2004. Stable-isotope probing of microorganisms thriving at thermodynamic limits:syntrophic propionate oxidation in flooded soil. Appl. Environ.Microbiol. 70:5778–5786.

35. Ma, K., R. Conrad, and Y. Lu. 2012. Responses of methanogen mcrAgenes and their transcripts to an alternate dry/wet cycle of paddy fieldsoil. Appl. Environ. Microbiol. 78:445–454.

36. Masters, C.I., J.A. Shallcross, and B.M. Mackey. 1994. Effect ofstress treatments on the detection of Listeria monocytogenes andenterotoxigenic Escherichia coli by the polymerase chain reaction. J.Appl. Bacteriol. 77:73–79.

37. Nishimura, S., T. Sawamoto, H. Akiyama. S. Sudo, and K. Yagi.2004. Methane and nitrous oxide emissions from a paddy field withJapanese conventional water management and fertilizer application.Glob. Biogeochem. Cycles 18:GB2017.

38. Noll, M., D. Matthies, P. Frenzel, M. Derakshani, and W. Liesack.2005. Succession of bacterial community structure and diversity in apaddy soil oxygen gradient. Environ. Microbiol. 7:382–395.

39. Nomura, M., R. Gourse, and G. Baughman. 1984. Regulation ofthe synthesis of ribosomes and ribosomal components. Annu. Rev.Biochem. 53:75–117.

40. Okabe, A., H. Oike, K. Toyota, and M. Kimura. 2000. Comparison ofphospholipid fatty acid composition in floodwater and plow layer soilduring the rice cultivation period in a Japanese paddy field. Soil Sci.Plant Nutr. 46:893–904.

41. Otsuka, S., I. Sudiana, A. Komori, K. Isobe, S. Deguchi, M.Nishiyama, H. Shimizu, and K. Senoo. 2008. Community structure ofsoil bacteria in tropical rainforest several years after fire. MicrobesEnviron. 23:49–56.

42. Plugge, C.M., W. Zhang, J.C. Scholten, and A.J.M. Stams. 2011.Metabolic flexibility of sulfate-reducing bacteria. Front. Microbiol.2:81.

43. Qiu, Y.L., S. Hanada, A. Ohashi, H. Harada, Y. Kamagata, and Y.Sekiguchi. 2008. Syntrophorhabdus aromaticivorans gen. nov., sp.nov., the first cultured anaerobe capable of degrading phenol toacetate in obligate syntrophic associations with a hydrogenotrophicmethanogen. Appl. Environ. Microbiol. 74:2051–2058.

44. Ramakrishnan, B., T. Lueders, P.F. Dunfield, R. Conrad, and M.W.Friedrich. 2001. Archaeal community structures in rice soils fromdifferent geographical regions before and after initiation of methaneproduction. FEMS Microbiol. Ecol. 37:175–186.

45. Roy, C., G. Talbot, E. Topp, C. Beaulieu, M.F. Palin, and D.I. Massé.2009. Bacterial community dynamics in an anaerobic plug-flow typebioreactor treating swine manure. Water Res. 43:21–32.

46. Sakai, S., H. Imachi, Y. Sekiguchi, A. Ohashi, H. Harada, and Y.Kamagata. 2007. Isolation of key methanogens for global methaneemission from rice paddy fields: a novel isolate affiliated with theclone cluster rice cluster I. Appl. Environ. Microbiol. 73:4326–4331.

47. Schloss, P.D., S.L. Westcott, T. Ryabin, et al. 2009. Introducingmothur: Open-source platform-independent, community-supportedsoftware for describing and comparing microbial communities. Appl.Environ. Microbiol. 75:7537–7541.

48. Selinger, D.W., R.M. Saxena, K.J. Cheung, G.M. Church, and C.Rosenow. 2003. Global RNA half-life analysis in Escherichia colireveals positional patterns of transcript degradation. Genome Res.13:216–223.

49. Spain, A.M., L.R. Krumholz, and M.S. Elshahed. 2009. Abundance,composition, diversity and novelty of soil Proteobacteria. ISME J.3:992–1000.

50. Spang, A., R. Hatzenpichler, C. Brochier-Armanet, T. Ratteri, P.Tischler, E. Spieck, W. Streit, D.A. Stahl, M. Wagner, and C.Schleper. 2010. Distinct gene set in two different lineages ofammonia-oxidizing archaea supports the phylum Thaumarchaeota.Trends Microbiol. 18:331–340.

51. Talbot, G., C.S. Roy, E. Topp, C. Beaulieu, M.F. Palin, and D.I.Massé. 2009. Multivariate statistical analyses of rDNA and rRNAfingerprint data to differentiate microbial communities in swinemanure. FEMS Microbiol. Ecol. 70:540–552.

52. Tamura, K., J. Dudley, M. Nei, and S. Kumar. 2007. MEGA4:Molecular evolutionary genetics analysis (MEGA) software version4.0. Mol. Biol. Evol. 24:1596–1599.

53. Tourna, M., M. Stieglmeier, A. Spang, et al. 2011. Nitrososphaeraviennensis, an ammonia oxidizing archaeon from soil. Proc. Natl.Acad. Sci. U.S.A. 108:8420–8425.

54. Urich, T., A. Lanzén, J. Qi, D.H. Huson, C. Schleper, and S.C.Schuster. 2008. Simultaneous assessment of soil microbial com-munity structure and function through analysis of the meta-transcriptome. PLoS One 3:e2527.

55. Wang, Q., G.M. Garrity, J.M. Tiedje, and J.R. Cole. 2007. NaïveBayesian classifier for rapid assignment of rRNA sequences into thenew bacterial taxonomy. Appl. Environ. Microbiol. 73:5261–5267.

56. Watanabe, T., M. Kimura, and S. Asakawa. 2007. Dynamics ofmethanogenic archaeal communities based on rRNA analysis andtheir relation to methanogenic activity in Japanese paddy field soils.Soil Biol. Biochem. 39:2877–2887.

57. Watanabe, T., V.R. Cahyani, J. Murase, E. Ishibashi, M. Kimura, andS. Asakawa. 2009. Methanogenic archaeal communities developed inpaddy fields in the Kojima Bay polder, estimated by denaturinggradient gel electrophoresis, real-time PCR and sequencing analyses.Soil Sci. Plant Nutr. 55:73–79.

58. Wind, T., S. Stubner, and R. Conrad. 1999. Sulfate-reducing bacteriain rice field soil and on rice roots. Syst. Appl. Microbiol. 22:269–279.

59. Yao, H., R. Conrad, R. Wassmann, and H.U. Neue. 1999. Effect ofsoil characteristics on sequential reduction and methane productionin sixteen rice paddy soils from China, the Philippines, and Italy.Biogeochemistry 47:269–295.